ADKINS, Judge.

Aggrieved when a jury awarded it only $4,000 instead of the $1.1 million it was seeking, Petitioner, Harford Sands, Inc. (Harford Sands) sought a new trial because a juror had obtained certain information from a source extraneous to the trial. The Circuit Court for Harford County rejected that plaint, as did the Court of Special Appeals. So do we. Balancing “the probability of prejudice from the face of the *139 extraneous matter in relation to the circumstances of the particular case,” Wernsing v. General Motors Corp., 298 Md. 406, 420, 470 A.2d 802, 809 (1984), we hold that the trial judge did not abuse his discretion when he denied Harford Sands’ motion for a new trial.

Harford Sands conducts a sand mining and processing operation in Harford County. The company owns 70 acres of land in Magnolia, Maryland, and on that land is a sand pit, approximately 25 acres in area, from which the sand to be processed is mined. Prior to 17 October 1983, there was a holding pond above the sand pit. The pond was designed to buffer the sand pit from storm water entering the property. An earthen dam separated the pond from the pit.

On 17 October 1983, Respondent Steven T. Groft and several other youngsters (hereinafter collectively Groft) entered the Harford Sands property and decided to construct a waterfall flowing from the holding pond into the sand pit. They dug a trench the width of the earthen dam so that water could flow from the pond into the sand pit. Eventually all of the water from the former drained into the latter. Employees of Harford Sands tried to repair the breach, but to no avail. The following day a rainstorm forced additional water into the pit. Within a week Harford Sands was able to pump the water from the sand pit, but another breach in the dam occurred shortly thereafter. As a consequence, the pit became a lake. The presence of this body of water made it unfeasible for Harford Sands to continue mining sand.

According to testimony produced on behalf of Harford Sands, it was difficult, if not impossible, for heavy equipment to reach the breach. Therefore, the breach could not immediately be repaired. Moreover, according to additional testimony, the cost of construction of an alternative system for keeping water out of the pit was great — beyond the financial grasp of the company. Thus the breach remained *140 unrepaired at the time of trial (March, 1988), leaving any sand in the pit unmined. Harford Sands (its witnesses said) had to procure sand for processing from another source, causing it additional financial loss.

There was conflicting testimony as to the accessibility of the breach to heavy equipment, the feasibility of repairing the breach, the cost of doing so, and the economic loss suffered by Harford Sands. Central to the case was the feasibility (financial and physical) of restoring the earthen dam, for if the breach could have been repaired quickly, at reasonable cost, much of the subsequent harm claimed by Harford Sands would not have occurred.

On the third day of the trial (18 March 1988) Richard Schafer, a storm water management expert, was testifying for Harford Sands on these matters. Sometime after the luncheon recess, the trial judge advised counsel that he had received a note from the jury posing this question: “In regards to repairing the breach, could you not use a concrete pumping machine to repair same from service road?” 1 Until that moment, there had been no mention of concrete pumping machines or possible use of one to overcome the claimed difficulties of access to the area of the breach. Counsel then asked Schafer whether a concrete pumping machine could have been used to repair the breach. Schafer thought not. In his view, the service road upon which the machine would have to be placed was “five hundred feet or something” from the breach. He opined that a pumping machine was “effective [only] three, maybe four, stories or thirty, forty, feet.” Moreover, such a machine could not have been used, he said, because the concrete pumped by it could not have been supported on the silty soil in the area of the earthen dam.

*141 At the end of Schafer’s testimony, trial counsel for Harford Sands requested, and the court gave, a cautionary instruction (supplementing one it had given at the outset of the trial). The court admonished the jury

that once you get this case for deliberation your decision is to be based solely on the evidence presented to you from both sides.

You are not to infuse your own ideas of any knowledge you may have of construction or storm water management or sediment control. It is only to be based upon the evidence that is presented ... through the various witnesses.

The trial proceeded. The outcome, as we have indicated, was a verdict in favor of Harford Sands for $4,000 in damages.

There was a motion for a new trial. In a memorandum in support of the motion Harford Sands alleged that a juror named Carroll O’Keefe had submitted the question about the concrete pumping machine and that he had thereafter, and in the face of the instruction just quoted, gone to a construction site near the county courthouse and investigated the nature and use of those machines. The memorandum further charged that on the basis of this extraneous information O’Keefe had concluded that Schafer’s testimony was unworthy of belief, had persuaded the other jurors to reach the same conclusion by explaining his “findings and conclusions about concrete pumping machines,” and had thereby managed to limit the damages to “the ridiculous sum of $4,000.00.”

At the hearing on the motion, Harford Sands produced O’Keefe and several affidavits. O’Keefe admitted that he had talked to workers at the construction site during the lunch break on 18 March, and that he had discussed with them the capabilities of concrete pumping machines. 2 He *142 admitted -authorship of the note about concrete pumps. He also said that “as we debated in the jury room, the concrete pump wasn’t even mentioned,” but the court correctly declined to consider that testimony.

The affidavits were from Andrew Scherer, Jeffrey Turner, Lee Saltzberg and Delores Morales. The first two affiants were construction workers who, they swore, were working at a construction site near the Harford County courthouse on 18 March 1988. Each declared that he had been approached by a man who identified himself as a juror in an ongoing trial in the circuit court, and who. inquired about the capabilities of concrete pumping machines. Scherer remembered telling the juror “that some concrete pumps could probably push concrete over lk mile laterally and as much as 8 to 10 stories high but. not at the same time.”

Saltzberg, who was trial counsel for Harford -Sands, asserted that he had spoken to O’Keefe after the- trial; that O’Keefe had admitted his conversations with the construction workers; that O’Keefe said he had, on the basis of those conversations, “decided that [Schafer’s] opinion regarding the cost of repairing the earthen dam was incorrect and his testimony was not to be believed”; and that O’Keefe had stated “that he was the strongest proponent of the defendants’ position on the jury,” Morales, who had *143 been a spectator at the trial, swore that she later discussed the trial with O’Keefe, who told her that he was the “ ‘strongest’ member of the jury”; that he knew (despite expert testimony to the contrary) that Harford Sands “could have repaired the breach in the dike right after it occurred” and that “he convinced the other jurors what the ‘real’ facts were by outlining all this information which was outside of the record in the ease, and seemed very proud that he had played this important role.”

The trial judge accepted the affidavits of Scherer and Turner, but not those of Saltzberg and Morales. As noted, he also refused to consider O’Keefe’s comments about what had or had not been discussed in the jury room. He accurately summarized the facts that were properly before him, and looking to Wernsing, correctly stated the relevant test as: “whether there is a probability of prejudice from the juror misconduct.” He recalled his several instructions to the jury on the subject of extraneous matters, as well as the evidence in the case. On the basis of all this, he could not

infer that any juror or that the jury as a whole ignored their obligations to decide the case based solely upon the evidence.

While it may be a fact that is established and I can consider that a juror talked to somebody about a pumping machine, I cannot say that juror based upon that limited inquiry, and I cannot infer absent some other fact that is before the Court, that the instructions of the Court were ignored.

He denied the motion. The Court of Special Appeals affirmed in an unreported opinion. Harford Sands, Inc. v. Groft, No. 1682, Sept. Term, 1988 (filed 10 August 1989).

II.

Before us, Harford Sands starkly depicts this case as one in which O’Keefe spun a “web of deceit” because he had a “predisposition against [Harford Sands’] theory of the case,” as evidenced by his submission of four hostile writ *144 ten questions, the climactic one being the concrete pump query. That last question, according to Harford Sands, was followed by O’Keefe’s flagrant violation of the court’s specific cautionary instructions (the conversations with the construction workers) and then by the juror’s taking the lead in convincing the other members of the panel that damages should be limited to “the ridiculous sum of $4,000.” The actual picture, however, is not bleak as Harford Sands pretends.

O’Keefe, of course, misbehaved when he talked to the workmen. “[I]t is highly improper for jurors to discuss with outsiders a case on trial before them....” Jos. F. Hughes v. Stockhausen, 212 Md. 559, 562,129 A.2d 844, 846 (1957). But the question before us is not whether or how O’Keefe should be punished for his misconduct. It is whether the circuit court judge abused his discretion when he denied Harford Sands a new trial. In deciding that issue, we are constrained to look at the record that was properly before that judge.

To begin with, there is nothing in the record that identifies O’Keefe as the author of any juror questions other than the one about the concrete pump. The trial judge made no findings in this regard, and we are not about to play handwriting experts in an effort to do so. Next, as we have already explained, O’Keefe did not seek out the construction workers in deliberate disregard of the court’s specific warning against extraneous material. He talked to the workmen before those instructions were given. See n. 2, supra, and accompanying text. Finally, there was nothing before the judge to show that O’Keefe had disclosed to his fellow jurors any information about concrete pumping machines or any views about the credibility of expert Schafer, or that he had played any part, leading or otherwise, in persuading the jury to bring in its $4,000 verdict. These conclusions, asserted by Harford Sands, are based on facts contained in affidavits properly excluded by the trial judge.

*145 The rule is well-settled in Maryland “that a juror cannot be heard to impeach his verdict.” Wernsing, 298 Md. at 411, 470 A.2d at 805; Oxtoby v. McGowan, 294 Md. 83, 101, 447 A.2d 860, 870 (1982). This rule applies “whether the jury conduct objected to be misbehavior or mistake.” Oxtoby, 294 Md. at 101, 447 A.2d at 870.

As Judge Rodowsky pointed out for the Court in Wernsing, in some other states, “the verdict impeachment prohibition is not as stringently applied as here.” 298 Md. at 415, 470 A.2d at 806. But the Maryland rule is stringent and of long standing. See Wernsing, 298 Md. at 411-412, 470 A.2d at 805, and cases there cited. The reasons for it are several. Evidence of jury deliberations would disclose secrets of the jury room and allow an opportunity for fraud and perjury. Brinsfield v. Howeth, 110 Md. 520, 530-531, 73 A. 289, 294 (1909). It would open the door to juror tampering. Id. Moreover, the rule seeks to avoid such risks as “harassment of jurors by disgruntled losing parties; removal of an element of finality from judicial decisions; and through allowing jurors to swear to alleged examples of reprehensible conduct, a decrease in public confidence in the judicial process.” Williams v. State, 204 Md. 55, 67, 102 A.2d 714, 720 (1954).

In Wernsing, a dictionary had been introduced into the jury room during jury deliberations. We allowed the testimony of a bailiff to the effect that the jury foreman had asked for a dictionary and that the bailiff had procured one and had handed it to the foreman. 298 Md. at 413-414, 470 A.2d at 806. The affidavits of Scherer and Turner are like the bailiffs testimony. They do not recount anything that happened in the jury room or the nature of the jury deliberations. They bear only on what one juror said and did outside the courthouse. See also Christ v. Wempe, 219 Md. 627, 642, 150 A.2d 918, 926 (1959). That also is true of O’Keefe’s statement that he conversed with the workmen during the lunch break. These evidentiary materials were properly considered by the trial judge.

*146 But O’Keefe’s remarks about what the jury discussed, and the affidavit of Morales and Saltzberg, were equally correctly rejected. They all related directly to juror deliberations and clearly suffered “the taint of possible post-verdict importuning.” Wernsing, 298 Md. at 413, 470 A.2d at 805. See also Oxtoby, 294 Md. at 100-101, 447 A.2d at 869-870.

Accordingly, we proceed to apply the teachings of Wemsing on the same factual basis as did the circuit court.

III.

In Wemsing this Court enunciated the rule for evaluating a motion for a new trial in a situation such as we have here. Judge Rodowsky, for the Court wrote

[w]here ... the precise extraneous matter is known but direct evidence as to its effect on the deliberations is not permitted, a sound balance is struck by a rule which looks to the probability of prejudice from the fact of the extraneous matter in relation to the circumstances of the particular case. It is the function of the trial judge when ruling on a motion for a new trial to evaluate the degree of probable prejudice and whether it justifies a new trial. That judgment will not be disturbed but for an abuse of discretion.

298 Md. at 419-420, 470 A.2d at 809. The trial court must balance the party’s “right to a fair trial with the policy prohibiting impeachment by a juror of the verdict.” Wernsing, 298 Md. at 419, 470 A.2d at 809. On appeal, the appellate court should defer to the trial court and overturn the lower court decision only “ ‘in those cases where there has been a plain abuse of discretion, resulting in palpable injustice.’ ” Stockhausen, 212 Md. at 563, 129 A.2d at 846 (quoting Rent-A-Car Co. v. Fire Ins. Co., 163 Md. 401, 409, 163 A. 702, 705 (1933)).

In Wernsing, the evidence was that a dictionary had been supplied to the jury and that the jury foreman considered a particular definition from the dictionary — a definition that *147 was inconsistent with the court’s instructions on proximate or legal cause. If the dictionary “approach to proximate cause were in fact applied by the jury, or even by [the foreman] alone, it would mean that all of the evidence, including expert opinions, ... would have been disregarded.” Wer nsing, 298 Md. at 420, 470 A.2d at 809. We held “[t]hat degree of probable prejudice [to be] so great that it was an abuse of discretion to deny a new trial.” Id.

Here all we know is that O’Keefe obtained extraneous information about the capabilities of concrete pumps that was inconsistent with Schafer’s testimony. We do not know that O’Keefe himself applied this information. We do not know that it was made available to any other member of the jury.

Nevertheless, Harford Sands insists that the trial judge abused his discretion in denying the new trial motion. It asserts, correctly, that in Wernsing, 298 Md. at 419, 470 A.2d at 808-809, we quoted from Palestroni v. Jacobs, 10 N.J.Super. 266, 271, 77 A.2d 183, 185 (1950), to the effect that:

“the test whether a new trial will be granted is whether the extraneous matter could have a tendency to influence the jury in arriving at its verdict in a manner inconsistent with the legal proofs and the court’s charge. If the extraneous matter has that tendency on the face of it, a new trial should be granted without further inquiry as to its actual effect.”

[Emphasis supplied]. Harford Sands avers that Wemsing adopted this test. And under this test, the argument runs, there should be a reversal, because O’Keefe’s information could have influenced him, and also could have been communicated to other jurors, and thus could have influenced them. That is, if there was a possibility of influence through the extraneous matter, a new trial must be ordered.

To adopt this rule would require a new trial in almost every case in which some extraneous matter has been *148 introduced into the jury room or somehow obtained by a juror, because almost anything is possible. That is not the rule we adopted in Wernsing. We cited Palestroni as illustrative of cases with records like the one there before us, in which jurors have described extraneous matter brought into the jury room, but have not been permitted to describe its effect on deliberations. We followed our reference to Palestroni with a discussion of State v. Ott, 111 Wis.2d 691, 696, 331 N.W.2d 629, 632 (1983), in which there was a remand for a new trial because “the probable effect [of extraneous information] upon a hypothetical average jury would be prejudicial.” [Emphasis supplied]. We then adopted the test we have quoted at the outset of this Part III, and it is couched in terms of “the probability of prejudice,” not the “possibility” thereof. That test is consistent with earlier decisions of this Court. See, e.g., Jos. F. Hughes and Rent-A-Car Co., both supra, and Safeway Trails, Inc. v. Smith, 222 Md. 206, 217-218, 159 A.2d 823, 829-830 (1960) (was extraneous matter of such a nature that its effect must fairly be held to have deprived injured party of fair trial?).

We bear in mind that we balance “the probability of prejudice from the fact of the extraneous matter in relation to the circumstances of the particular case.” Wernsing, 298 Md. at 420, 470 A.2d at 809. We recall, too, that the trial judge had thrice instructed the jury to consider only evidence in the case; we do not presume that this advice was disregarded. See Ferrell v. State, 318 Md. 235, 251, 567 A.2d 937, 945 (“ ‘It must be assumed that the jury ... logically and properly applied the instructions of the court’ ”) (quoting United States v. Flowers, 255 F.Supp. 485, 487 (E.D.N.C.1966)).

It is certainly true that O’Keefe, had he credited his extraneous information about concrete pumps, might have discredited all of Schafer’s testimony because of Schafer’s contradictory views of how far the machines could pump. Schafer was an important witness for Harford Sands on damages, and if one or more jurors rejected his testimony, *149 it would have had a serious effect on the plaintiffs case. But Schafer also testified that soil conditions would have made the use of a concrete pump impracticable. If the fact finder believed that, the distance such a machine could pump concrete would be unimportant.

Moreover, despite Harford Sands asserting that there was overwhelming and virtually uncontradicted evidence that it suffered great damages, that is not correct. Larry Stancill, the president of Harford Sands, was the chief witness for the corporation. The major harms of which he complained were (1) the flooding of the sand pit, thereby preventing the corporation from mining additional sand from it and forcing it to transport sand from elsewhere, and (2) the cost of repairing the breach or taking other measures to prevent continued flooding of the pit. But several pieces of testimony disputed these assertions. There was cross-examination of Stancill which, if believed, could have led to the conclusion that the sand in the pit had been virtually exhausted before the flooding. Several witnesses cast doubt on the extent of Harford Sands’ economic losses. There was testimony which, if believed, could have led to the conclusion that a second breach in the dike — one arguably not created by the defendants — was the cause of most of the flooding. There also was evidence which, if accepted, would have cast considerable doubt on Staneill’s credibility.

Most important, a deputy sheriff who had investigated the youths’ excavation shortly after it occurred said that Stancill, who had discussed with him the difficulty of repairing the original breach, also said that it would cost only about $4,000 to do so. While Stancill tried to explain this statement away, the jury might not have accepted his explanation. Patently, had the breach been promptly repaired for $4,000, a figure concededly within the means of Harford Sands, the other damages of which Harford Sands complained would not have occurred. When the jury announced its verdict the forewoman said: “We ... find the Defendants negligent without malicious intent and award to the Plaintiff the amount of $4,000.00 damage for the origi *150 nal breach.” That suggests that the jury believed the original breach could have been promptly repaired at modest cost, and had that been done, Harford Sands’ troubles would have been over.

While it is possible that this conclusion might have been the result of the improper influence of extraneous material, on this record, we cannot say that the concrete pump information probably resulted in prejudice to Harford Sands. In short, we see no palpable injustice here. The trial judge did not abuse his discretion when he denied the motion for a new trial.

JUDGMENT OF THE COURT OF SPECIAL APPEALS AFFIRMED. COSTS TO BE PAID BY PETITIONER.

. According to Schafer, a concrete pumping machine is "used to reach places that ... are very hard to get to." It is “like a hydraulic pump [that forces concrete] through a discharge hose, and it ends up at the point where you want it." He explained that "[i]t's commonly used for flooring on second or third floors or just hard-to-reach areas."

. Although at the motion hearing O’Keefe was somewhat confused about the order of events on 18 March (specifically, about whether his conversation had taken place before or after specific cautionary *142 instructions about avoiding extraneous information), the trial transcript makes it clear that the relevant sequence was this: (1) .court convenes at 10:35 a.m.; (2) Richard Schafer testifies — direct examination; (3) cross-examination of Schafer begins; (4) court recesses for lunch — 12:25 p.m. — 1:45 p.m.; (5) cross-examination resumes and is concluded; (6) court receives jury note about concrete pumping machine and reads it to counsel; (7) redirect examination of Scháfer, who is asked about concrete pumping machine; (8) counsel for Harford Sands requests and court gives cautionary instruction, advising jury not to “infuse your own ideas of any knowledge you may have of construction or storm water management” and to base its decision "upon the evidence that is presented ... through the various witnesses.” Since O’Keefe said (and no one denied) that he had talked to the construction workers during the luncheon recess on 18 March,, he must have had that conversation before the noté was submitted and before the court gave the special instructions.